White et al.
/
Chemi- and Bioluminescence of Firefly Luciferin
3199
Chemi- and Bioluminescence of Firefly Luciferin Emil H. White,* Mark G. Steinmetz, Jeffrey D. Miano, Peter D. Wildes, and Robert Morland Contribution from the Department of Chemistry, The Johns Hopkins University, Baltimore, Maryland 21 218. Receiued September 4, 1979
Abstract: Chemiluminescence of the ethoxyvinyl ester of 5,5-dimethylluciferin (3b) with base and I8O2 gave carbon dioxide in 39% yield enriched with 57-83% of 1 8 0 . The corresponding experiment with luciferin (3a) gave carbon dioxide in 20% yield with 76% I8O incorporation. The incorporation substantiates a dioxetanone mechanism for the light-producing pathway. The products from 5,5-dimethylluciferin ester 3b were 5,5-dimethyloxyluciferin (5b, 30%), dimethylluciferin (lb, 12%), and an isopropylidene rearrangement compound, 6 (44%).For firefly luciferin ester (3a), the products were oxyluciferin (5a, 30%) and dehydroluciferin (7, 73%). Excited-state yields of the oxyluciferins produced ranged from 33 to 53%. In contrast to the above results with “active” esters, the methyl ester of luciferin (8) on oxidation did not produce light and the major product formed was 4-hydroxyluciferin methyl ester (IO). The formation of dehydroluciferin is discussed.
Bioluminescence in the firefly is a complex phenomenon based on a simple set of chemical reactions of luciferin (la),
la, R = H
b, R = CH, 0 ATP Mg2+ luciferase
H 2
I
R
R=H b, R = CH,
3a,
0-0
R 4a,b O8
I+.
-hV
R 5a.b
0002-7863/80/ 1502-3 199$01.OO/O
ATP, oxygen, and magnesium ion catalyzed by the enzyme luciferase. Two steps are involved in this chemical production of excited states: ( I ) mixed anhydride formation to yield luciferyl A M P and pyrophosphate’ and ( 2 ) oxidation of the luciferyl adenylate2 (eq 1). Luciferyl A M P is also chemiluminescent on reaction with a base and oxygen, and other simpler “esters” are also a ~ t i v eWork . ~ with the latter compounds led to the reaction mechanism for the chemiluminescence of luciferin given in eq 2,435which was also proposed to hold for the biolumine~cence;~ preliminary results of our ’*Ostudy of the chemiluminescence were consistent with the mechanism.6 The chemically produced excited states were proposed to arise from a dioxetanone (4). This view was based on earlier studies of a common type of chemiluminescence in which oxygen is required a t some stage and in which dioxetanes are critical precursors of excited states.’ In the meantime, five contributions by De Luca et al. have appeared purporting to rule out the dioxetanone mechanism for both chemi- and bioluminescence.8 Oxygen isotopes were used and the results appeared to indicate that none of the oxygen atoms of the product carbon dioxide stemmed from the oxygen consumed. These conclusions have been widely cited.g Several defects in these studies have been pointed out which vitiate the mechanistic claim^.^^,^ The chief difficulties are (1) the chain of evidence was not completed in the work; that is, quantum yields and chemical yields, e.g., were assumed to be the same as values measured under nonidentical conditions in other laboratories, and (2) no proof was provided that the authors could recover the microliter quantities of carbon dioxide which had been produced from the reaction media. In the chemiluminescence experiments, a t least, it appears that adventitious carbon dioxide was analyzed throughout, since the claim was made that carbon dioxide produced in the reaction of luciferyl esters with potassium tert-butoxide was pumped directly out of the reaction mixture.I0 This is an impossible act in view of the high rate of reaction of tert-butoxide ion with carbon dioxide in solution and the great stability of the alkali metal salts of monoalkyl carbonates. I n our own studies no detectable carbon dioxide was released from such basic solutions. In contrast to the DeLuca conclusions, Johnson and Shimomura have shown in a study’’ of firefly bioluminescence that the carbon dioxide formed contains one atom of oxygen derived from the oxygen consumed. Recently, DeLuca et al.I2 have reinvestigated their earlier work and now conclude that the carbon dioxide formed in firefly bioluminescence does indeed contain an atom of oxygen derived from the oxygen. Thus, the general dioxetane-dioxetanone mechanism (eq 1 and 2) for both the bio- and chemiluminescence of firefly luciferin appears to be secure.
0 1980 American Chemical Society
3 200
Journal of the American Chemical Society
/
102.9
/ April 23, 1980
Table 1. Products from the Chemiluminescence of Luciferin Ethoxvvinvl Ester 3aa,h ~
run
solvent
3 4 (5) 6 (7Id
Me2SO Me2SO Me2SO TH F MezSO
I 2
Concentration of ester was 5-8 cmployed.
base/ ester
base
PhOK
72 3
t-BuOK
t-BuOK t-BuOK PhOK X
~~
02,
18 18
67
mm
5a, %
7, %
co2, 9c
228' 760 760 760 209
34 27 30 49 (42) 16 (17)
46 54 73 20 (22) 51 (55)
22(17)
Values of duplicate runs in parentheses. ' Air-saturated solution. Vacuum-line techniques
M.
Table 11. Dependence of Chemiluminescence Quantum Yields for Luciferin Ethoxyvinyl Ester (3a) on Oxygen Pressure 02,
mm
9ci 1 .o
760 374 217 I15 65
8
I .o
1 .o
COzCH,
0.70 0.35
Quantum yield relative to that at 760 m m
1. O,, PhOK 2.
H '
02.
9
In the present paper we give full details of our I 8 0 study of the 5J-dimethyl derivative of luciferin,13 and also an account of parallel studies of the parent luciferin.
Results Products. Carboxylic acid derivatives of luciferin with good leaving groups produce a brilliant chemiluminescence on treatment with a base and oxygen. This reaction in the case of the ethoxyvinyl ester of 5,5-dimethylluciferin (3b) leads to 5,5-dimethyloxyluciferin (5b), a rearrangement product ( 6 ) , 5,S-dimethylluciferin (lb), carbon dioxide (eq 3), and a small amount of an unknown s ~ b s t a n c e . ~ ~ ~ ' ~
I
1. 0,.PhOK MeSO
44% 6
+ lb +
CO2
1%
39%
+ hr
(3)
The chemiluminescent reaction of the ethoxyvinyl ester of the parent luciferin (3a) with base and oxygen produced oxyluciferin (5a),I5dehydroluciferin (7),16carbon dioxide, and an unidentified degradation product (eq 4 and Table I); in-
10
was produced; the products were the methyl ester of dehydroluciferin (9, 17% yield) and the methyl ester of 4-hydroxyluciferin (10,48%yield) (eq 5). The major product (10) was characterized by its spectral data and transformation to dehydroluciferin methyl ester (9) by acid-catalyzed dehydration. Dehydroluciferin methyl ester (9) was not formed from hydroxy ester 10 under the reaction conditions employed for ester 8 oxidation, nor during workup and product isolation. Chemiluminescence. The chemiluminescence maximum for the ethoxyvinyl ester of 5,5-dimethylluciferin (3b) was found to be 63 1 nm with potassium phenoxide, potassium tert-butoxide, or potassium imidazolate as the base in dimethyl sulfoxide solutions saturated with oxygen. The wavelength was found to be independent of base type or c ~ n c e n t r a t i o nThe .~ quantum efficiency, @-L, with a tenfold excess of base was 0.09 for dilute solutions ( M).4,6a3'4 The chemiluminescence spectrum of the ethoxyvinyl ester of firefly luciferin (3a) was similar to that of other luciferin derivati~es,~ exhibiting a, , A, of 625 nm (red light) in dimethyl sulfoxide with 1.63 X M potassium phenoxide as the base ( 1 2.5 equiv). In tetrahydrofuran, an orange emission at 595 nm was observed under these conditions. For the ethoxyvinyl ester of firefly luciferin (3a), the ~PCL was 0.09 at 1 X M ester. The dependence of quantum yield on oxygen pressure was determined (Table 11) in order to optimize conditions for the I8O2 tracer study; for 5.71 X M solutions of ester 3a, decreased quantum yields were observed for oxygen pressures less than 150 nm. Fluorescence Measurements. 5,5-Dimethyloxyluciferin (5b) exhibited a long-wavelength absorption a t 580 nm in dimethyl sulfoxide containing an excess of potassium phenoxide (300 equiv), with fluorescence occurring at 633 nm. Similar results were reported for potassium imidazolate or triethylamine as the base.4 The absorption and fluorescence spectra of firefly oxyluciferin (5a) proved to be dependent on base concentration (Table 111). With a large excess of potassium phenoxide in dimethyl sulfoxide, the absorption, , A, of 5a was 475 nm, and excitation at this wavelength produced a single yellow-green fluorescence
- HoqHNr CO,H
3a
OH
1. 0,.base 2. Ht
5a
+
S
7
+
+
co* hv (4) terestingly, the product distribution appeared to be solvent dependent. The analogue of compound 6 was apparently not formed in the parent luciferin case. Firefly luciferin is stable under the reaction conditions and it is not the source of dehydroluciferin. Yields of carbon dioxide were obtained during the I8O tracer study using vacuum line techniques and are summarized in Table I (runs 6 and 7). In the absence of oxygen, the reaction of luciferin ester 3a with base was nonchemiluminescent; neither oxyluciferin (5a) nor dehydroluciferin (7) was produced, the hydrolysis product, luciferin, being the only product detected. The methyl ester of luciferin (8) was treated with potassium phenoxide and oxygen under similar conditions but no light
White et ai.
/ Chemi- and Bioluminescence of Firefly Luciferin
3201
Table 111. Fluorescence of Fireflv Oxvluciferin (Sa). Dehvdroluciferin (7).and Spent Reaction Mixtures
base/compd ratio
compd Sa a 5a a
absorption Xmax,
300 1.4
nm
475 497,570 (sh)
Sa a
0.27
375,497,570 (sh)
Sa spent chemiluminescence reaction mixtureh
0.0 2.0
375 472,570 (sh)
dehydroluciferin (7)
300
47 1
M in argon-saturated dimethyl sulfoxide. From 2.09 X a 9.03 X m L of dimethyl sulfoxide, flushed with argon for the measurements. Table IV. ISO Incorporation in the Carbon Dioxide Produced from Luciferin Esters 3a,ba-C % COz labeled
run 8 9
IO II
ester
base
with I8Od
3b 3b 3b 3a
t-BuOK t-BuOK t-BuOK PhOK
83 67 57 76
Ester and excess base in dimethyl a Vacuum-line techniques used. sulfoxide at 25 OC. 100%phosphoric acid for acidification except for run 1 1, which used phenol. Raw values corrected for the atom % ‘SOin the starting oxygen, but not for exchange.
emission a t 570 nm. At a base:5a ratio of 1.4:lthe absorption spectrum gave,,A 497 nm with a less intense shoulder a t 570 nm. Excitation into the short-wavelength band gave a fluorescence emission a t 570 nm with a shoulder a t 610 nm. Excitation into the weak longer wavelength shoulder gave only red emission (628 nm). With less than 1 equiv of base, an absorption at 375 nm was present (un-ionized oxyluciferin) along with the 497-and 570-nm bands; the fluorescence spectrum was similar to that obtained with 1.4 equiv of base. A spent ester 3a chemiluminescence reaction mixture also exhibited a 570-nm shoulder in the absorption spectrum; excitation into the shoulder gave red emission a t 628 nm paralleling the oxyluciferin (5a) result (Table 111). The major remaining bands were assigned to dehydroluciferin (Table 111). An additional 2.6 equiv of base caused disappearance of the 570-nm absorption and 628-nm emission, whereas the 555-nm emission of 7 remained unchanged. Acidification of the spent reaction mixture with glacial acetic acid gave a spectrum identical with that of dehydroluciferin (7)(A,,, 350 nm). Tracer Studies. Our product studyI4 of 5,5-dimethylluciferin ester (3b) chemiluminescence showed that oxygen was necessary for 5,5-dimethyloxyluciferin (5b) formation and also for light emission; the same result was obtained for luciferin ester 3a chemiluminescence. In the tracer experiments, base was added to a solution of a n ester of luciferin containing dissolved 1 8 0 2 . After light emission ceased, an acid was added to liberate the carbon dioxide from the carbonate salt which had been formed in the basic medium and the carbon dioxide was analyzed. In a regular apparatus sealed by rubber septa, the I8O incorporation was low (1-16%) in the carbon dioxide produced following acidification with strong acids even When precautions were taken to avoid traces of water and atmospheric carbon dioxide. The highest incorporation (22%) was obtained when anhydrous phosphoric acid” was used. In the same run 60% of the total 5,s-dimethyloxyluciferin was labeled with 180.I n a control run, labeled C02 was carried through one acidification cycle with the result that 62% of the I8O was “washed out”. Thus, 58% of the C02 from the luciferin would have been labeled
fluorescence
hex, nm 475 497 570 497 570 375 472 570 47 1
Xem,
nm
570 570,610 (sh) 628 570,610 (sh) 628 428 555 628 552
mmol of ester l b and 2 equiv of potassium phenoxide in 3.5
prior to acidification, corresponding to the percent of the 5,5-dimethyloxyluciferinthat was labeled. In order to minimize the problem of exchange by adventitious water,’* vacuum-line techniques were used in which all manipulations such as purification, transfer of reagents, mixing, and initiation of chemiluminescence were performed in a closed system, The yield of isotopic enrichment in the carbon dioxide increased dramatically under these conditions to 83% in one case, and it averaged 72% (runs 8-10,Table IV). Comparable enrichment in I8O was found in the isolated 5,5-dimethyloxyluciferin (94% I 8 0 ) . In one run, H2I80 was added to the reaction mixture and 1 6 0 2 was used as the oxidant. Serious exchange occurred to give carbon dioxide containing 52% C i 6 0 1 8 0and 34% C180180, but some unlabeled C 0 2 remained ( 14%).18b Carbon dioxide obtained from the chemiluminescence of the ethoxyvinyl ester of the parent firefly luciferin (3a) had an isotopic enrichment of 76% I8O (Table IV). In this case a large excess of phenol relative to potassium phenoxide was used to buffer the medium so that carbon dioxide could be collected without resorting to strong acid to decompose the intermediate potassium phenyl carbonate. Improvement over the 100% phosphoric acid method used in ester 3b runs was not realized, however. Discussion The data will be discussed in terms of the general mechanism for firefly luminescence proposed earlier (Chart I).4 The ethoxyvinyl esters of the luciferins (3) were used in this study since they could be readily prepared analytically pure; the adenylate analogue (2) has not been reported as an isolable, pure solid and the phenyl ester is difficult to prepare.I9 The Light Emitter in Luciferin Chemi- and Bioluminescence. Yellow-green light (ca. 560 nm) is normally produced by fireflies, but under certain conditions (e.g., a t higher temperatures) red light (ca. 630 nm) is emitted.*Oa The in vitro luciferase-catalyzed luciferin oxidation also yields yellow-green light normally, but red-light emission occurs a t low pHs, at higher temperatures, and in the presence of heavy metals2 The chemiluminescence of the ethoxyvinyl ester of firefly luciferin la, on the other hand, yields only red’light under a variety of conditions.2’ Whether red or yellow-green light is emitted depends on whether ionization of 12 to 13 can compete with emission (Chart I ) . In the chemiluminescence of ethoxyvinyl ester 3a, external base does not compete and only red emission is observed. In the luciferase-catalyzed reaction, a basic group on the enzyme does compete, and yellow-green emission is normally seen. At low pHs, red light is produced presumably because a basic group on the enzyme is being protonated. When that basic group is alkylated, it has been found that only red-light emission occurs irrespective of the pH of the medium.22 In the case of 5,5-dimethylluciferin, the anion of the oxy-
3202
Journal of the American Chemical Society
Chart I. Firefly Luciferin Luminescence
\ loo%, carbon dioxide must be a product of the light-producing pathway. The similarity of the oxyluciferin (30%) and carbon dioxide (39%) yields is consistent with the proposed reaction (eq 1). On the average, 70% of the carbon dioxide molecules formed in the vacuum line runs contained an atom of ISO(Table 1V). Since a minimum of 33% excited states was formed in the oxidative pathway, the data is a t least consistent with the view that I8O is incorporated in the carbon dioxide during the generation of excited states. In the run with the least washing out of the label, the I8Oincorporation was 83%; since 83 33 > loo%, an overlap occurs and excited states must have been formed with incorporation. The products formed, their distribution, the spectral data, and the labeling data are all consistent with the view that dioxetanone intermediates are involved in the chemiluminescence of firefly luciferin (Chart I);24 the bioluminescence of firefly luciferin also follows this pathway.' As an alternative to the reaction pathway illustrated in the upper pathway of Chart I, an elimination of a type well known for acid chlorides might occur to give a ketene (11). T h e formation of dioxetanones from ketenes and oxygen has been reported.25 High yields of singlet states are obtained in the chemi- and bioluminescence of firefly luciferin. Yet simple dioxetanones and dioxetanes decompose to give principally triplet states. In an effort to explain the high singlet yields, Koo, Schmidt, and Schuster26(eq 7), and also M ~ C a p r (eq a ~ 8), ~ have proposed electron-transfer mechanisms for the generation of excited states from dioxetanones of luciferin;2s reaction mechanisms were not specified. As a point of interest, the dioxetanone ring in 4 defines a plane perpendicular to the plane of the conjugated benzothiazole and thiazoline rings. To be analogous to the intermolecular electron transfer processes proposed for the chemiluminescence of dioxetanones plus fluorescent compounds,26 the formation of 19 would require through-space electron transfer between the mutually perpendicular rings of 4. Some analogy for this type of transfer exists in the work of
19
L
-
20
+
12
CO,
(8)
Simmons and Fukunaga31 on electron transfer in spiro systems. Dehydroluciferin (7). The title compound has been detected in fireflies, it can be isolated from that source,32and it has been prepared by the air oxidation of luciferin in aqueous base.16 W e now find that the methyl ester of dehydroluciferin, 9, is formed in the air oxidation of the methyl ester of luciferin; no light is produced in these oxidations. The oxidation probably proceeds by the pathways shown in eq 9. 8-
PhOK
-&yy
CO,CH,
'SJ 21
OOH I
+
+
1
' . r t \HH
Y 10
22
A second product of the oxidation is the methyl ester of 4hydroxyluciferin (IO). This compound may stem from a reduction of the hydroperoxide (22), possibly by the carbanion precursor 21. Dioxetanone formation would be slow in this case because of the poor leaving-group ability of the methoxide ion, and the hydroperoxide (22) thus meets a different fate. Corresponding alcohol analogues of hydroperoxides have been observed in the chemiluminescence of Cypridina luciferin,33 l ~ p h i n eindoles,35 ,~~ and t r y p t ~ p h a n s . ~ ~ The chemiluminescent reaction of the ethoxyvinyl ester of luciferin 3a with oxygen in the presence of base yields much more dehydroluciferin (7) (as the free acid after workup, 48-73%) than does the methyl ester (1 7% yield as the ester), suggesting that the carboxyl group is involved in the excess production of dehydroluciferin. Possible sources of the dehydroluciferin are outlined in structures 4' and 23 (followed by 0-
I
fi -base 4'
23
reduction or hydrolysis of the peracid formed). Perepoxides have been proposed as intermediates in the addition of oxygen to ketenes,25 the addition of singlet oxygen to hindered alk e n e ~ , ~and ' in certain e p o ~ i d a t i o n s . ~ ~ The ready formation of dehydroluciferin in the chemiluminescence of the ethoxyvinyl ester suggests that it might be formed in vivo39 in an analogous way from luciferyl adenylate. Thus, the dehydroluciferin found in fireflies may be essential
3204
Journal ofthe American Chemical Society
102.9
1 April 23, 1980
rather than artifactual; possibly it is involved as an inhibitor in shaping the flash nature of the light emission in fireflies. 2,40 Experimental Section
follow: luciferin, 11.5 min; dehydroluciferin, 1 I .5 min; oxyluciferin, 18.7 min; luciferin ethoxyvinyl ester, 21 .0 min; anthracene. 3 I .S min. The identitj of each peak in product mixtures was tentatively determined by comparison of the ultraviolet spectrum at stopped flow with authentic samples and also by comparison of retention time. The Introduction. Melting points were determined with a Thomasidentities were confirmed (when noted below) by isolation using Hoover capillary melting point apparatus and are uncorrected. Elepreparative LC; fractions were freeze-dried and analyzed by mass mental analyses were performed by Galbraith Laboratories, Knoxville, spectroscopy. Tenn. N M R spectra were measured usinga J E O L C O M H - l 0 0 a n d Products of Luciferin Ethoxyvinyl Ester (3a) Chemiluminescence. peaks are reported in parts per million relative to tetramethylsilane. mmol) of luciferin T o a stirred solution of 1.25 mg (3.58 X Mass spectra were obtained with an Hitachi Perkin-Elmer RMU-6D ethoxyvinyl ester in 5 mL of dry, air-saturated dimethyl sulfoxide spectrometer at 70 eV. Fluorescence and chemiluminescence spectra (distilled from sodium hydride) was added a 0.50-mL (0.26 mmol) were determined on a Hitachi Perkin-Elmer Model MPF-2A specaliquot of 0.51 M potassium phenoxide in 1 : 1 phenol-dimethyl sulftrofluorimeter. High-pressure liquid chromatography was carried out oxide (v/v) via a syringe. When the red chemiluminescence ceased, with a Waters Associates M6000A pump, a U6K injector, and a 0.2 mL of acetic acid was added. The mixture was concentrated to ca. Beckman LC-25 variable-wavelength detector. 1 mL at 45 “ C (0.5 mm) using a liquid nitrogen cooled cold finger. Ethoxyvinyl 2-(6’-hydroxy-2’-benzothiazolyl)-5,5-dimethylMethylene chloride was added and the suspension was washed with A2-thiazoline-4-carboxylate(3b) (5,5-dimethylluciferin ethoxyvinyl water and then with saturated sodium chloride to remove salts: the solution was then dried over anhjdrous sodium sulfate and evaporated ester) and 2-(6’-hydroxy-2’-benzothiazolyl)-4-isopropylideneA*-thiazolin-5-one (6)were prepared as described by White, Suzuki, to dryness at 0.5 mm. The residue was dissolved in methanol and analyzed by LC using column B (vide supra). Assaj gave oxylu~iferin~’ and Miano.I4 The following compounds were prepared by established procedures: 2-(6’-hydroxy-2’-benzothiazolyl)-A~-thiazoline-5,5-di-(1.23 X I 0-3 mmol, 34%), dehydroluciferinIh (1.63 X IOp3 mmol, methyl-4-one (5b) (5,5-dimethyloxyl~ciferin),~~ 2-(6’-hydroxy-2’46%). and an unidentified product at 22.5 niin retention time [A,,, benzothiazolyl)-5,5-dimethyl-A2-thiazoline-4-carboxylic acid ( l b ) (stopped flow) 354 nm]. Each component was isolated preparatively ( 5 , 5 - d i m e t h y l I u ~ i f e r i n ) , ~2-(6’-hydroxy-2’-benzothiazoly1)-A2’ b j LC (vide supra) and the mass spectra were determined: m/r (re1 thiazoline-4-carboxylic acid ( l a ) (firefly l ~ c i f e r i n ) , ’and ~ 2-(6’intensity) for oxyluciferin 250 ( I 4), 176 ( 1 00): for dehydroluciferin hydroxy-2’-benzothiazolyl)thiazole-4-carboxylicacid (7) (dehydro278 (loo), 234 (62), 194 (46). Examination o f t h e M 2 ( m / e 280) luciferin).Ih 2-(6’-Hydroxy-2’-benzothiazolyl)-A2-thiazolin~4-one peak of dehydroluciferin showed that luciferin was absent. Since luciferin and dehydroluciferin each had identical retention time by LC, (5a) (oxyluciferin) was provided by N . Suzuki.Ij Ethoxyvinyl 2-(6‘-Hydroxy-2’-benzothiazolyl)-AZ-thiazoline- the dehydroluciferin isolated preparatively was chromatographed on cellulose, eluting with 1 : 1 methanol-w,ater. Luciferin was abyent at 4-carboxylate (Luciferin Ethoxyvinyl Ester 3a). T o a solution of 198 Rf 0.8 and only one spot of dehydroluciferin was present at Rf 0.5 mg (0.707 mmol) of luciferinIh in 20 mL of dry tetrahydrofuran (distilled from lithium aluminum hydride) under argon was added (blue fluorescent). Additional runs a r e given in Table I (runs 2-5), I .OO mL ( 1 1.4 mmol) of ethoxyacetylene4* followed by 0.70 m L of where L C assay was conducted with t h e crude reaction mixture 0.03 1 M mercuric acetate (0.022 mmol) in dry tetrahydrofuran. After without workup. I9 h, T L C (silica gel, 1 : 1 ethyl acetate-benzene) showed essentially A similar run in the absence of oxygen (carried out under high complete conversion of luciferin (Rf 0.0) to ester product (Ry 0.55). vacuum) was nonchemiluminescent. LC assay (column A ) showed only luciferin (100%) and no oxyluciferin. T L C on cellulose with 70% LC assay using column B (vide infra) gave a ratio of 93:7 for ester and luciferin, respectively. The reaction mixture was poured into ether, ethanol and 30% 1 M ammonium acetate eluent showed only luciferin at Ry 0.76 and no dehydroluciferin at Rf 0.50. washed with saturated sodium bicarbonate and then water, dried over Methyl 2-(6‘-Hydroxy-2‘-benzothiazolyl)-Az-thiazoline-4-caranhydrous sodium sulfate, and concentrated in vacuo to give 68.9 mg boxylate (Luciferin Methyl Ester, 8).To a solution of 173 mg (0.61 8 (28%) of the luciferin ethoxyvinyl ester as a pale yellow solid, mp > 176 OC dec (with gas evolution): IR (KBr) 3.25 (br), 5.68,6.00,6.37, 6.76, mmol) of luciferin in 25 mL of tetrahydrofuran at 0 OC was added 3 mL of ca. 0.3 M ethereal diazomethane with stirring. T L C on silica 8.20, 8.70 p : N M R (Me2SO-d6) 6 1.24 ( t , J = 6 Hz, 3 H , CH3), gel eluting with 50% ethyl acetate in benzene (v/v) showjed rapid 3.7-4.2 (m,6 H , CH2), 5.67 ( t , J = 8 Hz. I H , C H ) , 7.08 ( d d , J = 3, I 1 Hz, I H, aromatic), 7.43 (d, J = 3 Hz, 1 H , aromatic), 7.94 (d, J conversion of luciferin (R, 0.0) to R j 0.41 yelloh, fluorescent methyl ester. After 3 min. the reaction mixture was concentrated to dryness = I 1 Hz, I H. aromatic), 10.2 (s, I H , O H ) ; addition of deuterium at 0.05 mm to obtain 175 mg of a tan solid, m p 190-192 OC dec. oxide caused exchange of the phenolic proton at 6 10.2 and partial Crystallization from an acetone-cyclohexane mixture gave 124 mg exchange (ca. 75% loss of signal) of the 4 proton at 6 5.67; U V A,,, ( T H F ) 332 n m ( E 17 900), 272 (8220). (68%) of a colorless, crystalline solid, m p 192.5-193.0 OC dec: IR (KBr) 3.12 (br), 5.76, 6.38, 6.74, 6.87, 8.14 p ; N M R (acetone-dh) Anal. Calcd for CljHl4NZ04S2: C , 51.42; H, 4.03; N , 7.99. Found: C,51.16:H,4.08;N,7.89. 6 3.86 ( d , J = 9 Hz, 2 H , CH2), 3.87 ( s , 3 H , -CO>CH3), 5.56 (t. J Higher yields were obtained with ethyl acetate as the extraction = 9 Hz, 1 H, C H ) . 7.32 (dd. J = 2, 9 Hz. 1 H, aromatic). 7.67 (d, J = 2 Hz, 1 H , aromatic), 8.1 3 (d, J = 9 Hz. 1 H, aromatic); U V A,;, solvent. After drjing, the solution was concentrated to 2 mL at 0.05 ( M e O H ) 332 nm ( E 1 5 700), 269 (7400); M S (70 eV) nile (re1 i n m m and 4 niL of cyclohexane was added. Cooling, suction filtration, and drying at 0.05 m m gave a 53% yield of pale yellow, crystalline ester tensity) parent 294 (20), 235 (loo), 176 ( 16).J4 Anal. Calcd for C12Hlol\j,03S2: C, 48.96; H, 3.42; N, 9.52. Found: ~ ~ h i uc ahs pure b j T L C ( R ~ 0 . 5 2yellow , fluorescent); only a very faint C , 49.19; H , 3.54: N , 9.36. spot representing luciferin was seen at R j 0.0. Solutions of the ester in dichloromethane could be washed with sodium bicarbonate soluMethyl 2-(6‘-Hydroxy-2‘-benzothiazolyl~thiazole-4-carboxylate (DehydroluciferinMethyl Ester, 9). The method was adapted from that tions with no detectable hydrolysis. Chromatographic Procedures for Products Derived from Luciferin of White et A solution of 90.0 mg (0.557 mmol) of 6-hydroxybenzothiazole-2-thiocarboxamideJb and 423 mg (2.33 mmol) of Ethoxyrinyl Ester Chemiluminescence. Product assays involved addition of a known amount of anthracene internal standard followed methyl b r o ~ n o p y r u v a t ei n~ ~I O m L of methanol was stirred for 21 h by high-pressure liquid chromatography (LC) using the following at 25 “C. T L C on silica gel with 50% ethyl acetate in benzene showed columns: column A, 30 cm X 3.9 m m Waters Associates Bondapak partial conversion of the thiocarboxamide [R, 0.50 (yellow fluorescent)] to dehydroluciferin methyl ester [Rj 0.41 (blue fluorescent)]. Clx eluting nith 50% acetonitrile in water ( v / v ) at 0.5 mL/min flow rate: column B, a combination of column A in series with 1 5 cm X 3.9 Refluxing for 1.3 h under nitrogen yielded a suspension of a pale yelloh. crystalline solid which was then filtered and washed with mni Whatman Partisil-I0 S C X with 557cacetonitrile in water (v/v) eluent at 0.5 mL/min flow rate. The ultraviolet variable wavelength methanol to give 126 nig ( 7 7 6 ) of the ester, mp 288-290 OC dec. dctector was calibrated at 380 nm for the response of each component Recrystallization from tetrahydrofuran in methanol gave a yellowish relative to anthracene using known mixtures. Luciferin and luciferin solid, mp 289-291 O C dec. IR (KBr) 3.25 (br), 5.82, 6.27,6.36,6.62, ethoxyvinyl ester were not detectable at 380 nm at the concentrations 6.96. 7.45, 7.82. 8.18 p ; N M R (acetone-dh-Me2SO-dh) 6 4.02 (s, 3 employed in chemiluminescence r u n s below: these compounds were H. C02CH3), 7.30 (dd, J = 2, 9 Hz, 1 H, aromatic), 7.69 (d, J = 2 ’~ assayed instead at 334 nm. The retention times follow: l ~ c i f e r i n ,15.3 Hz, 1 H , aromatic), 8.1 I (dd, J = 2 , 9 HI., 1 H, aromatic), 8.84 (s, 1 min; dehydroluciferin,lh 15.5 min; o x y l u ~ i f e r i n , ’18.0 ~ ~ ~min; ~ anH, vinyl), and 10.15 (5, I H, phenolic); U V A,,, ( M e O H ) 351 nm ( t thracene, 58.5 min using column A. For column B, the retention times 22 000), 275 ( I O 400); M S (70 eV) ni/e (re1 intensity) 292 ( l o o ) , 261
+
White et al.
/ Chemi- and Bioluminescence of Firefly Luciferin
(22), 234 (40), 194 (16). Anal. Calcd for C ~ ~ H ~ N ~CO , 49.30; ~ S ZH:, 2.76; N , 9.58. Found: C , 49.08: H , 2.87; h,9.46. This compound could be prepared also from reaction of dehydroluciferin with excess diazomethane i n tetrahydrofuran at 0 OC followed by concentration to dryness at 0.05 mm. Oxidation of the Methyl Ester of Luciferin (8).T o an oxygen-saturated solution of 57.3 mg (0.195 mmol) of ester in 50 mL of dry tetrahydrofuran (distilled from sodium) was added ca. 0.42 mmol of a solid potassium phenoxide-phenol complex.46 N o chemiluminescence was observed. After the mixture was stirred for 3 min, T L C on silica gel with 5070 ethyl acetate in benzene (v/v) eluent showed two components: Rf 0.39, blue fluorescence (dehydroluciferin methyl ester), and R j 0.18, yellow fluorescence (4-hydroxyluciferin methyl ester). No oxyluciferin at Rf 0.62 (blue fluorescence) was observed. The reaction mixture was acidified with 2 drops of concentrated hydrochloric acid and concentrated to ca. 5 mL i n vacuo. Ethyl acetate was added and the solution was washed with saturated solutions of sodium sulfite, sodium bicarbonate, and sodium chloride. After the solution was dried over anhydrous sodium sulfate, the T L C was unchanged from the above. The extract was concentrated in vacuo, silica gel was added. and the slurry was dried at 0.05 mm. The solid was poured onto a 20 X 2 cm silica gel column (Will, Grade 922) packed in benzene. The chromatography was monitored by T L C and UV. The following results were obtained taking 25-mL fractions: fractions 1-2, benzene, nil; fractions 3-7, 10%ethyl acetate in benzene, ca. 0.10 mg o f blue fluorescence material (unidentified); fractions 8-9,20% ethyl acetate in benzene, nil; fractions 10-1 3, 20% ethyl acetate in benzene, 9.6 mg ( I 7% yield) of dehydroluciferin methyl ester: fraction 14, 20% ethyl acetate in benzene, 0.09 mg of a mixture of dehydroluciferin methyl ester and 4-hydroxyluciferin methyl ester; fractions 15-27, 40% ethyl acetate in benzene, 38 mg o f 4-hydroxyluciferin methyl ester. The dehydroluciferin methyl ester was pure by T L C and was identified by UV, mass spectrum, and mixture melting point with an authentic sample. The 4-hydroxyluciferin methyl ester in benzene-ethyl acetate was decolorized using Norite and the solution was concentrated to dryness. The residue was dissolved in a small amount o f tetrahydrofuran. Addition of benzene induced precipitation to afford 14 mg of a pale yellow solid, mp 205-206 OC dec. A second crop was obtained of I5 mg, mp 204-206 OC dec: total yield, 29 mg (48%); IR (KBr) 3.03 (br), 5.76, 6.35. 6.72, 8.13 p ; N M R (MezSO-ds) 6 3.61 (d, J = 12 Hz, 1 H, CHz), 3.92 (s. 3 H , COfZH3),4.19 ( d , J = 12 Hz, 1 H, CH2),7.34 (dd, J = 2 , 9 Hz, aromatic), 7.68 (d, 1 H , J = 2 Hz, aromatic), 8.17 (d. 1 H, J = 9 Hz, aromatic), 10.5 (s, I H, phenolic O H ) ; UV, ,A, ( M e O H ) 337 nm ( E 16 SOO), 269 (6830); MS (70 eV) m/e (re1 i n tensity) parent 310 (8), 292 (32). 251 ( I O O ) , 234 (9), 207 (8), 177 (66), I76 (34). The 4-hydrox>luciferin methyl ester was further characterized by dehydration to dehydroluciferin methyl ester. A solution of 2.95 mg (9.52 X nimol) o f hydroxy ester 10 in 5 mL of methanol with I drop of concentrated sulfuric acid was refluxed for I O min. T L C on silica gel with 50% ethyl acetate in benzene showed complete conversion from the yellow fluorescent hydroxy ester ( R , 0.15) to the blue fluorescent dehydroluciferin methyl ester ( Rf 0.40). The LV spectrum had changed from A,,; 337 nm to ,A, 352 nm, and the change in optical density showed 80% conversion. Upon cooling to room temperature. I .3 mg (46%) of dehydroluciferin methyl ester was obtained as a pale yellow. crystalline solid. mp 189-1 91 OC. A similar oxidation of the methyl ester of luciferin was carried out with dry dimethyl sulfoxide (distilled from sodium hydride) as the solvent. The yellow solid residue obtained after workup and concentration to dryness was chromatographed on a 20 cm X 20 cm X 0.25 mm silica gel plate (Eastman), eluting with 50% ethyl acetate in benzene (v/v). After two passes, the top band (blue-white fluorescent) gave dehydroluciferin methyl ester ( 1 2% by UV assay). The yellow band gave 38% of 4-hydroxyluciferin methyl ester (by UV assay). Stability of 4-Hydroxyluciferin Methyl Ester ( I O ) to Oxidation. The procedure followed that for luciferin methyl ester oxidation in tetrahydrofuran (vide supra). A solution of 1.68 mg (5.42 X IO-) mmol) of 4-hydroxyluciferin methyl ester i n I mL of oxygen-saturated tetrahydrofuran containing 0.024 mmol of potassium phenoxide was stirred for I O min and acidified with I drop of concentrated hydrochloric acid. T L C on silica gel with 50% ethyl acetate in benzene (v/v) showed the yellow fluorescent starting material at Rj 0.18 and a faint yellow fluorescent material at the origin. Dehydroluciferin methyl
3205
ester was not present at R/0.40; 5 X mmol (0.99") would have been detected based on T L C of a known mixture. The residue after workup was chromatographed on a 8 X 1 cm silica gel (Will, Grade 922) column eluting with 30% ethyl acetate in benzene. The starting 4-hydroxyluciferin methyl ester was obtained in the second 20" fraction. After concentration to dryness, LV assay in methanol showed 2.72 X IO-) mmol (50.2% recovery). Stability of Luciferin Methyl Ester (8)to Base without Oxygen. All operations were performed on a vacuum line in order to rigorously exclude oxygen. A solution of 12.2 mg (0.0437 mmol) of luciferin methyl ester in 15 mL of tetrahydrofuran was degassed by two freeze-pump-thaw cycles. Concurrently 3 mL (0. I7 mmol) of 0.056 M potassium phenoxide in tetrahydrofuran was degassed in an adjoining vessel. A side arm contained 0. I mL of frozen (77 K ) glacial acetic acid for later use. The base solution was tipped into the reaction flask, and the green-colored reaction mixture was stirred for I O min. The acetic acid was then thawed and washed into the reaction mixture to give a colorless solution. T L C on silica gel eluting with 50% ethyl acetate in benzene showed only yellow-green fluorescent luciferin at 332 nm was identical with that methyl ester a t Rf0.35. The,,A, of the starting ester; the absorbance indicated a recovery of 9 I%. Stability of Luciferin to Potassium Phenoxide and Oxygen. The procedure was identical with that for luciferin ethoxyvinyl ester chemiluminescent runs and employed I .05 nig (3.75 X IO-) mmol) of luciferin and 0.152 mmol of potassium phenoxide i n 4.0 mL of oxygen-saturated dimethyl sulfoxide. The reaction was monitored by T L C on cellulose with a n eluent of 70% ethanol 30% 1 M ammonium acetate. Over a period of I h, only luciferin a t Rf 0.67 was observed. Dehydroluciferin at RJ 0.46 was absent. Preparative T I L on a 20 cm X 20 cm X 0.5 mm cellulose plate eluting with 50% methanol in water gave recovered luciferin a t Rf 0.75. A very faint blue fluorescent band at Rf0.5 gave 1.7 X mmol (0.4%)ofdehydroluciferin by UV analysis. Phenyl 2-(6'-Trifluoroacetoxy-2'-benzothiazolyl)-Az-thiazoline4-carboxylate. To a stirred mixture of 65 mg (0.23 mmol) of luciferin and 240 mg (2.5 mmol) of phenol was added 3 mL of trifluoroacetic anhydride at 0 "C under nitrogen. The reaction flask was allowed to reach room temperature. and then it was kept at that temperature for 1 h. The excess phenol and trifluoroacetic anhydride were removed under vacuum to give a yellow solid, mp 117-1 19 OC: IR (KBr) 5.56, 5.78 p; N M R (CDC13) 6 4.48 (d, J = 5 Hz, 2 H , CHz), 6.20 (t, J = 5 Hz, 1 H, C H ) , 7.91 (s, 5 H , phenyl), 8.08 (dd, J = 2, 8 Hz, 1 H, aromatic), 8.47 (d, J = 2 Hz, I H, aromatic), 8.80 (d, J = 8 Hz, 1 H , aromatic); MS m/e 452 (parent), 331 (-CO?Ph); U V (MezSO) 335, 273 nm; T L C (silica gel, eluting with 50% ethyl acetate in benzene) R/ 0.70 (major component), 0.55 (possibly phenyl ester), 0.0 (minor; luciferin and/or dehydroluciferin). General Procedure for Chemiluminescence Quantum Yield Measurements. The apparatus used to determine light yields for luciferin ethoxyvinyl ester consisted of a K446 photomultiplier, a Fluke Model 41 2 B power source at 600 V, and a Keithley electrometer to measure total charge (proportional to the total light emitted). The photomultiplier was corrected for its differing response vs. ~ a v e l e n g t h . The ~' chemiluminescence spectrum of luciferin ethoxyvinyl ester in dimethyl 625 nm and sulfoxide using potassium phenoxide as base gave A,;, was similar to that reported for other luciferin ester derivatives (vide infra).4b Quantum yields were obtained relative to luminol ( @ C L = 0.01 1 ), a procedure used to obtain chemiluminescence quantum yields for 5,5-dimethylluciferin ethoxyvinyl e ~ t e r . ~ , ' ~ Potassium Imidazolate. Potassium amide in liquid ammonia was prepared from 9.2 g (0.24 g-atom) of potassium in 500 mL of liquid ammonia under nitrogen with a catalytic amount of ferric nitrate nonahydrate. Imidazole ( I 7 g, 0.25 mol) (decolorized with Norite in hot benzene followed by crystallization) was added in portions over 15 min to 0.24 mol of potassium amide in 500 mL of liquid ammonia. The ammonia was evaporated under a stream of nitrogen and the remaining solid was pulverized, washed with ether, and dried under vacuum to give an off-white powder. Effect of Oxygen Partial Pressure on the Chemiluminescence Quantum Yield of Luciferin Ethoxytinyl Ester (3a). Each run employed 3.40 X IO-' mmol of luciferin ethoxyvinyl ester in 5.2 mL of dimethyl sulfoxide (degassed by three freeze-pump-thaw cycles). The sample solution (5.6 mL total volume) for each determination was contained in a 26.4-niL cylindrical Pyrex cell: a 0,024-in. pinhole aperture was interposed between the sample cell and photomultiplier (in these runs, a 1P21 photomultiplier was used). Oxygen was introduced into the
+
3206
Journal of the American Chemical Society
evacuated cell and the pressure measured. Light was produced by adding 0.4 mL (0.14 mmol) of 0.36 M potassium phenoxide (prepared from potassium tert-butoxide and excess phenol) in nitrogen-saturated dimethyl sulfoxide (v/v), Chemiluminescence Spectra. The chemiluminescence of the ethoxyvinyl ester of luciferin normally was too fast to measure with the scanning monochromator of the fluorimeter. Thus, spectra were obtained by a point-by-point plot of intensity vs. wavelength; each point was determined by adding an aliquot of potassium phenoxide by syringe to the ester and obtaining the intensity of emission at a particular wavelength. An 8.44 X M dimethyl sulfoxide solution gave A,, 625 f 3 nm; in tetrahydrofuran the maximum was 595 f 3 nm. In two cases, the chemiluminescent reaction was slow enough to be measured by a scanning monochromator: addition of a pellet of potassium hyM dimethyl sulfoxide solution of ester gave A,, droxide to a I X 625 f 3 nm. Addition of 2 molar equiv of potassium tert-butoxide to a 7.03 X I O v 5 M solution of ester in tetrahydrofuran gave A,, 595 & 3 nm. Fluorescence Measurements. Corrected spectra were obtained by standard proced~res.~8-50 The @" was determined for 5,5-dimethyloxyluciferin (5b) in dimethyl sulfoxide with excess potassium imidazolate at high dilution4x ( A = 0.050 at 540 nm). T h e corrected fluorescence spectrum (A,, 635 nm) was compared to the corrected fluorescence spectrum (A,, 580 nm, @n = 0.6948) of a dilute rhodamine B standard in ethanol having an identical optical density of 0.050 a t 540 nm. Two determinations gave @n = 0.6 I , 0.64 (average 0.62). As a check, the fluorescence quantum yield of rhodamine B (A540 = 0.050) was determined against quinine bisulfate in I N sulfuric acid as a standard (A,, 470 nm, = 0.5548). The excitation wavelength for the standard was 363 nm ( A = 0.055). Two determinations gave @f = 0.69 and 0.7 I , in agreement with the literature. Fluorescence studies with oxyluciferin (5a) were more qualitative in nature and the spectra were not corrected. Oxyluciferin (5a) was 5.03 X M in argon-saturated dimethyl sulfoxide. For each determination, a known amount of potassium phenoxide in dimethyl sulfoxide was added. After each fluorescence spectrum, the absorption spectrum was obtained. The sample was then acidified with glacial acetic acid and the absorption spectrum redetermined; the absorption spectrum was unchanged from that prior to base addition. The volumes of base and glacial acetic acid used changed the total volume of the sample solution by only 1% at most. The results are summarized in Table 111. Calculation of Percent Carbon Dioxide Labeled with '*O Using Mass Spectroscopic Data. Accurate peak intensities were obtained by slow pen recorder scans of carbon dioxide samples. All peak heights were corrected for background. The atom percent, A, of I8Oin carbon dioxide was obtained from A=-
2(R
+ I)
x 100
where R = 46/44 (m/e value^).^' The percentage N of the total carbon dioxide labeled with one lSOwas then given by N =
2(A - 0.20) x 100 5 - 0.20
where 0.2 is the natural abundance of lXOand 5 is the atom percent I8O in the I8Oused in the chemiluminescent reactions. Vacuum Line Isotope Runs with Luciferin Ethoxyvinyl Ester (3a). An all-glass cell with two side attachments was used and the cell was attached to the vacuum line through a series of three traps (numbered from the reaction cell). The first attachment was used for the distillation of MezSO (from sodium hydride) into the reaction cell. The sccond attachment allowed the sublimation of potassium tert-butoxide into a sample of phenol contained therein. After evacuation the dimethyl sulfoxide and phenol were degassed by two freeze-pump-thaw cycles. After the potassium tert-butoxide was sublimed further into the side arm and the attachment was resealed with a torch, dimethyl sulfoxide was distilled into this attachment until ca. 1 mL was collected; 6 mL was then distilled into the main part of the cell and cooled to -78 OC. Oxygen which had been cooled to 77 K for 2 h prior to reaction was introduced into the reaction vessel with slight warming to attain ca. 200 mm pressure. After the contents was thawed in the apparatus, the base was dissolved in the 1 mL of Me2SO and the solution was then tipped into the reaction pot to initiate red chemiluminescence. After 0.5 min, the reaction mixture was frozen to - 196
1 102.9 1 April 23, 1980
"C. With trap 1 at -107 "C (isooctane-Nl) and trap 2 at -196 OC, the oxygen was pumped out of the system. The reaction mixture was thawed and the volatiles were distilled into the trap system for 18 min while pumping. The trap system was isolated from the apparatus and the rest of the vacuum line. Trap 2 was isolated from trap 1 and its contents thawed and recondensed into the bottom tip of the trap using liquid nitrogen. Trap 2 was then warmed to -107 OC with frozen isooctane and the volatiles were distilled into trap 3 cooled to - 196 "C. T r a p 2 was isolated and trap 3 warmed to room temperature. From the known volume of the system and the pressure, the yield of carbon dioxide was determined. The carbon dioxide was then transferred to a bulb for mass-spectral analysis. The above procedure successfully eliminated volatile impurities such as ethyl acetate and terr-butyl alcohol. After the traps were cleared, acetic acid (0.2 mL) was distilled through the system into the reaction pot and the above collection procedure was performed again. Additional carbon dioxide was not obtained after acidification. The reaction mixture was then analyzed by LC using column A (vide supra). The I8O used (Miles Laboratories) was 92.82 atom 9" enriched. Run 6: 3.14 X mmol of luciferin ethoxyvinyl ester; ca. 0.21 mmol of potassium phenoxide; 209 mm l6O2 (Linde extra dry); 0.393 m m (22.0%) carbon dioxide (natural abundance); oxyluciferin, 4.94 X mmol (15.7%); dehydroluciferin, 1.61 X 10-3 mmol ( 5 I ,370). Run 7: 3.05 X mmol of luciferin ethoxyvinyl ester; ca. 0.27 mmol of potassium phenoxide; 207 m m I8O2; 0.298 m m (17.3%) carbon dioxide, R = 0.10, 9.3% I8O in carbon dioxide; oxyluciferin, 5.12 X mmol (16.8%); dehydroluciferin, 1.67 X mmol (54.8%). Run 11: 3. I2 X IO-) mmol of luciferin ethoxyvinyl ester; ca. 0.23 mmol of potassium phenoxide; 204 m m l8O2; carbon dioxide not quantitated, R = 2.42, 76% I8Oin carbon dioxide; oxyluciferin, 4.37 X mmol (14.0%); dehydroluciferin, 1.15 X mmol (3 6.9%). Control Run. No luciferin ethoxyvinyl ester, ca. 0.21 mmol of potassium phenoxide. 156 mm I8O2; 0.010 m m volatile material unidentified; no volatiles obtained after acidification. Standard Isotope Runs with 5,5-Dimethylluciferin Ethoxyvinyl Ester (3b). The apparatus consisted of a 5-mL flask with a rubber septum sealed side arm; this was connected through two U-tube traps in series to a high-vacuum line. A solution of 2.04 mg (5.40 X mmol) of 5.5-dimethylluciferin ethoxyvinyl ester in 2.0 mL of dry dimethyl sulfoxide was degassed by two freeze-pump-thaw cycles. One atmosphere of oxygen-18 (91.6 atom % excess) was introduced. A 10 molar excess of potassium phenoxide in dimethyl sulfoxide (prepared from sublimed potassium tert-butoxide and 1.5 equiv of phenol) was added by syringe to give red chemiluminescence. After 5 min, the reaction vessel was evacuated and the mixture neutralized with 100% phosphoric acid'' in dimethyl sulfoxide. Vigorous evolution of carbon dioxide was observed. With trap 1 at -78 OC and trap 2 at -196 O C , the reaction vessel was evacuated. After trap 2 was isolated and warmed to -78 O C , the volatiles were collected and analyzed by mass spectroscopy. The spectrum showed a ratio of m/e 46/m/e 44 of 0.27. The percentage of carbon dioxide labeled with I8Owas thus 22%. The 5,5-dimethyloxyluciferin produced was isolated by preparative T L C as described p r e v i o ~ s l y .The ' ~ mass spectrum showjed tn/e 278 (77-nim peak height) and 280 (104 mm) to give 60% incorporation of IxOlabel (see section below for method of calculation.) Isotopic Exchange in Carbon Dioxide. A solution of 7.6 mg (0.068 nimol) of freshly sublimed potassium fert-butoxide in 2.0 mL of dimethyl sulfoxide was degassed by two freeze-pump-thaw cycles. Carbon dioxide (0.035 mm), 22% of which contained an atom of I8O, uas condensed into the reaction vessel cooled to -196 OC. After thawing, the reaction vessel was evacuated and 100% phosphoric acid" in dimethyl sulfoxide was added by syringe. With trap I at -78 O C and trap 2 at - 196 OC, 0.030 mm of carbon dioxide was collected in trap 2. The mass spectrum showed that only 8.5% of the carbon dioxide was labeled with lag. Vacuum Line Isotope Runs with 5,5-Dimethylluciferin Ethoxyvinyl Ester (3bi. The apparatus was similar to that used for the parent luciferin ester except that it contained an additional side arm to contain phosphoric acid. For each run a glass boat containing 5,5-dimethylluciferin ethoxyvinyl ester was introduced into the main chamber of the flame-dried apparatus under nitrogen. Potassium terr-butoxide (ca. I O mg) was placed in the base arm; all openings at this stage were sealed with septa; the apparatus was partially evacuated by means of a syringe, and then the base arm was sealed with a torch. In the same
White et al. / Chemi- and Bioluminescence of Firefly Luciferin manner, 100% phosphoric acid" was introduced t o a n a c i d side a r m which was then sealed off, A 5-mL aliquot o f d i m e t h y l sulfoxide containing sodium hydride (heated p a r t i a l l y t o f o r m dinisyl anion) was filtered through glass wool into the side a r m pot (through a sleeve) which was then sealed o f f w i t h a torch while frozen. T h e apparatus was evacuated t o 2 p. With the d i m e t h y l sulfoxide at -78 "C, potassium tert-butoxide was sublimed t o the upper part o f the base side a r m at ca. 170 "C b y heating w i t h an oil bath; the unsublimed residue was removed b y c u t t i n g o f f the end o f the tube w i t h a torch. A f t e r the dim e t h y l sulfoxide was degassed b y three freeze-pump-thaw cycles, 3 m L was distilled i n t o the reaction pot (cooled t o - I 9 6 "C) w h i l e pumping. T h e side a r m pot was then removed w i t h a torch w h i l e keeping the apparatus evacuated. O x y g e n - I 8 (53.5 a t o m % excess) was introduced. Once thawed, the ester solution was used t o dissolve the base t o give red chemiluminescence. A f t e r evacuation o f the vessel and acidification, the volatiles were passed i n t o t r a p 1 (-78 "C) and t r a p 2 (-196 "C). A f t e r t r a p 2 was isolated f r o m t r a p 1 a n d warmed t o -78 "C, the carbon dioxide was quantitated using the ideal gas law. This was collected in a bulb cooled t o -196 "C for mass-spectral analysis. T h e results are summarized below. In run I , 5,5-dimethyloxyluciferin was isolated b y preparative cellulose TLC as described p r e v i o ~ s l y .T~h~e mass spectrum showed m/e 278 (83-mm peak height) a n d 280 (91 mm) thus giving a calculated l8O incorporation o f 94%. T h e data are summarized as follows: m m o l luciferin ethoxyvinyl ester reactant, pressure (% yield) o f carbon dioxide, r a t i o R = m/e 46/m/e 44, calculated percent o f I8O in carbon dioxide. Run 8: 6.65 X m m o l ester, 0.055 mm ( 5 5 % ) , R = 0.81, 83% I8O in carbon dioxide. Run 9: 1.48 X mmol ester, 0.055 m m (28%), R = 0.57,67% I8O in carbon dioxide. Run 1 0 I .46 X IO-* m m o l ester, 0.065 mm (34%), R = 0.45, 57% I8O in carbon dioxide. I n this run m/e 46 peak height required a 10% reduction due t o a contribution b y contaminating d i m e t h y l sulfide; the ratio m/e 62/m/e 46 o f 0.36 (authentic material) was used t o make the correction. Control Run. No ester, 0.01 mm o f volatile material, impure carbon dioxide m/e 48 (37%), 46 (loo%), 44 (74%). T h e a t o m percent incorporation in carbon dioxide was calculated51bto be 42%, which gives I.55 atoms o f 1 8 0 per carbon dioxide after correction for the a t o m percent excess in the l8O2 used. Analysis of [180]5,5-Dimethyloxyluciferin (5b). Mass spectra o f both labeled a n d unlabeled 5,5-dimethyloxyluciferin were run using the solid inlet system o f the mass spectrometer. T h e spectrum o f the unlabeled sample had a molecular ion (P) at m/e 278, a P I peak w i t h a n intensity 15.5% t h a t o f P, and a P 2 peak w i t h a n intensity 9.3% that o f P. The height o f the labeled 5,5-dimethyloxyluciferinmolecular ion at m/e 280 was corrected for the P 2 contribution from unlabeled 5,5-dimethyloxyluciferin by m u l t i p l y i n g the observed m/e 278 peak height b y 0.093 and subtracting this value from the observed m/e 280 peak. These values were then analyzed b y the following method: x i s the observed peak height o f m/e 280 corrected for the P 2 peak of m/e 278, y i s the peak height o f m/e 278, and the values o f x and y corrected for the a t o m percent oxygen-I 8 used are x' and y', respectively, where x' = x/B a n d y ' = y (x' x) and B i s the atom percent oxygen-18 in the oxygen-I8 used. T h u s the percentage o f the total 5.5-dimethyloxyluciferin labeled w i t h one atom o f oxygen-I8 is equal t o x'/(x' y') 100. Isotope Run with l8O Water and I6O Oxygen. T h e procedure for the lXOisotope run was as described for the I8O molecular oxygen r u n except for the following modifications. A f t e r the 5.5-dimethylluciferin ethoxyvinyl ester ( II.8 mg, 3.12 X I O - 2 m m o l ) was pushed i n t o the central reaction chamber, I8O water (I00 pL, 8 1.8 a t o m % I8O)was placed into a small piece o f glass tubing which had one end sealed. The tube filled w i t h I8O water was pushed u p the base side a r m i n t o the central reaction chamber. A t this point the central reaction chamber was frozen i n a dry ice-acetone b a t h until the vessel was evacuated. I t was then cooled to l i q u i d nitrogen temperatures for the sublimation o f potassium tert-butoxide and the distillation step. Enough dimethyl sulfoxide was distilled t o make the final solution ca. 1.5-2.0 M in I8O water. Matheson Research Grade oxygen (analyzed < I p p m COz) was used instead of oxygen-18. T h e carbon dioxide collected (195 p ) was analyzed b y mass spectroscopy. T h e heights o f the m/e 44, 46, a n d 48 peaks in the mass spectrum were 66, 136, and 8 4 mm, respectively. Analysis a n d correction o f these data for the 81.8 a t o m 70 I8O used gave t h e following percentages o f labeled carbon dioxide: ci8~'*02. = 34.2%; C's-1602, 51.8%; Cl6-I6O2, 14.0%.
+
+
+
+
-
+
-
3201
Acknowledgments. We thank the U S . Public Health Service for i t s financial support (Grant G M 19488) and Martha Umbreit and Dr. N. Suzuki for valued assistance. References and Notes (1) W. C. Rhodes and W. D. McElroy, J. Biol. Chem., 233, 1528 (1958). (2) W. D. McElroy, H. H. Seliger, and E. H. White, Photochem. Phofobiol., I O , 153 (1969). (3) T. A. Hopkins, H. H. Seliger, E. H. White, and M. W. Cass, J. Am. Chem. SOC.,89, 7148 (1967). (4) (a) E. H. White, E. Rapaport, H. H. Seliger, and T. A. Hopkins, J. Am. Chem. Soc., 91, 2178 (1969); (b) E. H. White, E. Rapaport, H. H. Seliger, and T. A. Hopkins. Bioorg. Chem., 1, 92 (1971). (5) F. McCapra. Y. C. Chang, and V. P. Francois, J. Chem. Soc., Chem. Commun., 22 (1968). (6) (a) E. H. White, J. D. Miano, and M. Umbreit, J. Am. Chem. Soc., 97, 198 (1975); (b) M. J. Cormier. D. M. Hercules, and J. Lee, Eds.. "Chemiluminescence and Bioluminescence", Plenum Press, New York, 1973, pp 358-359. (7) E. H. White and M. J. C. Harding, Photochem. Photobiol., 4, 1129 (1965); K. R. Kopecky, J. H. van de Sande. and C. Mumford. Can. J. Chem. 46,25 (1968). (8) (a) M. DeLuca and M. E. Dempsey, Biochem. Biophys. Res. Commun., 40, 117 (1970); (b) M. DeLuca and M. E. Dempsey in ref 6b, p 345; (c) M. Deluca, Adv. Enzymol., 44,37 (1976); (d) M. DeLuca, M. E. Dempsey, K. Hori, and M. J. Cormier, Biochem. Biophys. Res. Commun., 69,262 (1976); (e) F. I. Tsuji, M. DeLuca, P. D. Boyer, S. Endo, and M. Akutagawa, ibid., 74, 606 (1977). (9) (a) M. M. Rauhut, "Kirk-Othmer: Encyclopedia of Chemical Technology", Vol. 5, 3rd ed., Wiley, New York, 1979, pp 416-450; (b) R. C. Hart and M. J. Cormier, Photochem. Photobiol., 29, 209 (1979);(c) J. P. Henry and A. M. Michelson, ibid., 27, 855 (1978); (d) J. W. Hastings and T. Wilson, ibid., 23, 461 (1976); (e) J. Lee, ibid., 20, 535 (1974);(f) F. McCapra, Acc. Chem. Res.. 9, 201 (1976); (9) M. J. Cormier, J. Lee, and J. E. Wampler, Annu. Rev. Biochem., 44, 255 (1975); (h) M. J. Cormier, J. E. Wampier, and K. Hori, Fortschr. Chem. Org. Naturst., 30, 1 (1973): (i)F. McCapra, Endeavor, 32, 139 (1973). (10) References 8b,d and private communication from M. DeLuca. Carbon dioxide had also been "collected" after bubbling it through a solution of potassium tert-butoxide in tert-butyl alcohol (M. DeLuca, M. E. Dempsey, K. Hori, J. E. Wampler, and M. J. Cormier, Proc. Natl. Acad. Sci. U.S.A., 68, 1658-1660(1971). (1 1) 0. Shimomura, T. Goto, and F. H. Johnson, Proc. Natl. Acad. Sci. U.S.A., 74, 2799 (1977). (12) J. Wannlund, M. DeLuca, K. Stempel, and P. D. Boyer, Biochem. Biophys. Res. Commun., 81, 987 (1978). (13) Reference 6a is a preliminary communication on this subject. (14) E. H. White, N. Suzuki, and J. D. Miano, J. Org. Chem., 43, 2366 (1978). (15) (a) N. Suzuki, M. Sato, K. Okada, and T. Goto, Tetrahedron Left, 4683 (1969); (b) N. Suzuki and T. Goto, Agric. Biol. Chem., 36, 2213 (1972); (c) N. Suzuki, M. Sato, K. Okada, and T. Goto. Tetrahedron, 28, 4065 (1972); (d) N. Suzuki and T. Goto, ibid., 28, 4075 (1972); (e) T. Goto. I. Kubota. N. Suzuki, and Y Kishi in ref 6a, p 325 (16) E H White, F McCapra. and G F Field, J Am Chem SOC, 85, 337 119631 (17) Prepa;ed from 85% phosphoric acid and excess phosphorus pentoxide (L. F. Fieser and M. Fieser, "Reagents for Organic Synthesis", Wiley, New York, 1967). (18) (a) The haif-life for exchange in water is ca. 1000 sat pH 5; see also G. A. Mills and H. C. Urey, J. Am. Chem. Soc., 62, 1019 (1940); 61, 534 (1939); B. H. Gibbons and J. T. Edsall, J. Biol. Chem., 238, 3502 (1963); E. Ho and J. M. Sturtevant, ibid., 238, 3499 (1963). (b) In a similar experiment with luciferyl adenylate, DeLuca et al. report only that 0.65 atom of oxygen was incorporated.8d (19) In the synthesis of the phenyl ester from luciferin, phenol, and trifluoroacetic anhydride, the Otrifluoroacetyl derivative of the phenyl ester is obtained initially; subsequent removal of the trifluoroacetyl ester group selectively proved difficult. In the analogous reaction of 5,5-dimethylluciferin, considerable rearrangement product 6 is formed concurrently14 (see ref 5). (20) E. N. Harvey, "Bioluminescence", Academic Press, New York, 1952: (a) p 428; (b) p 445. (21) We have not observed yellow-green light emission at high base/ester ratios, reaction conditions that Seliger and Hopkins report led to yellow-green emission in the case of the phenyl ester.4 (22) E. H. Whiteand B. Branchini, J. Am. Chem. Soc., 97, 1243-1245 (1975); Methods Enzymol., 46, 537 (1977). (23) H. H. Seliger and W. D. McEiroy, Arch. Biochem. Biophys., 88, 136 (1960). (24) Of course, as the reaction system approaches totaldryness, the oxygen atoms of the oxygen gas used would of necessity end up in the carbon dioxide by any one of a number of mechanisms. The present work was ated to counter the claim that none of the oxygens of the carbon dioxide produced stemmed from the reactant o x y g m 8 In the systems we used, water was present as shown by the exchanges occurring and thus the extent of incorporation of l8ois of significance. (25) (a) N. J. Turro. M-F. Chow, and Y. Ito, J. Am. Chem. Soc., 100,5580 (1978); (b) N. J. Turro, Y. lto, M-K. Chow, W. Adams, 0 . Rodriquez. and F. Yang. ibid., 99, 5836 (1977). (26) J-Y. Koo, S. P. Schmidt, and G. B. Schuster, Proc. Natl. Acad. Sci. U.S.A., 75, 30 (1978). (27) F. McCapra, J. Chem. Soc., Chem. Commun., 946 (1977). (28) The two groups cite negligible or nonchemi- and bioluminescence of the 6'-Omethyl ether of l ~ c i f e r i n * ~ in. support ~~ of the electron-transfer mechanism. The fluorescence efficiency of O-methyloxyluciferin has not been reported, however. In the event that the efficiency is low, the ob-
3208
Journal of the American Chemical Society
servations can be explained simply by the low fluorescence of the emitter. The claim that the "methylated ketone itself fluoresces efficiently'2~26is not supported by the reference cited.29 (29)F. McCapra, Pure Appl. Chem., 24, 611 (1970). (30)Reference 4b; cf. p 104. (31)H. E. Simmons and T. Fukunaga, J. Am. Chem. Soc., 89,5208(1967). (32)8. Bitler and W. D. McElroy, Arch. Biochem. Biophys., 72, 358 (1957). (33)0.Shimomura and F. H. Johnson in ref 6a, p 337. (34)E. H. White and M. J. C. Harding, Photochem. Photobiol., 4, 1129 (1965); the 2,5,5-triphenyI-4-imidazoiinone derives from an alcohol intermediate which rearranges. (35)(a) F. McCapra and Y. C. Chang, Chem. Commun., 522 (1966);(b) C. L. Stevens and R. J. Gasser, J. Am. Chem. Soc., 79,6057(1957). (36)M. Nakagawa and T. Hino, Acc. Chem. Res., I O , 346 (1977). (37)F. McCapra and I. Beheshti, J. Chem. Soc., Chem. Commun., 517
(1977). (38)P. D. Bartlett. Chem. Soc. Rev., 5, 149 (1976). (39)Of course, the yield of dehydroluciferin would be limited by the local quantum yield of bioluminescence: in vitro, the value of 0.88f 0.12has been reported.23
(40)J. L. Denburg, R. T. Lee, and W. D. McElroy, Arch. Biochem. Biophys., 134, 381 (1969):J. A. Nathanson, Science, 203,65-68 (1979). (41)E. H. White, H. Worther, G. F. Field, and W. D. McElroy, J. Org. Chem., 30,
/
/ April 23, 1980
102:9
2344 (1965). (42)E. R. H. Jones, G. Eglington, M. C. Whiting, and B. L. Shaw. "Organic Syntheses", Collect. Vol. IV, Wiley, New York, 1963,p 404. (43)Authentic oxyluciferin was provided by Dr. N. Suzuki. (44)The physical data listed here for this compound supersede values reported previo~siy.~~
(45)N. F. Blair and C. G. Stuckwisch, J. Org. Chem., 22, 82 (1957). (46)Prepared from potassium tert-butoxide and 2 equiv of phenol followed by concentration to dryness at 0.05 mm to give a solid of ca. 1:l mole ratio potassium phenoxide/phenol. (47)(a) J. Lee and H. H. Seliger, Photochem. Photobiol., 4, 1015 (1965);(b) J. Lee, A. S. Wesley, J. F. Ferguson, and H. H. Seliger in "Bioluminescence in Progress", F. H. Johnson and Y. Haneda. Eds., Princeton University Press, Princeton, N.J., 1966,p 35. (48)J. G.Caivert and J. N. Pitts, Jr., "Photochemistry", Wiley, New York, 1966, pp 799-804. (49)C. A. Parker, "Photoluminescence in Solutions", Elsevier, Amsterdam, 1968,p 253. (50)D. R. Roberts and E. H. White, J. Am. Chem. Soc., 92,4861 (1970). (51) (a) We did not observe m l e 48 to any extent greater than ca. 5% of d e 46 unless noted specifically. (b) When d e 44,46,and 48 were present in the carbon dioxide mass spectrum, the method of Shimomura and Johnson" was used.
Conformational Studies on Humulene by Means of Empirical Force Field Calculations. Role of Stable Conformers of Humulene in Biosynthetic and Chemical Reactions Haruhisa Shirahama,* Eiji Gsawa, and Takeshi Matsumoto Contribution from the Department of Chemistry, Faculty of Science, Hokkaido University, Sapporo, 060, Japan. Received September 17, 1979
Abstract: The conformational behavior of humulene, an 1 I-membered, carbocyclic, biosynthetically important sesquiterpene, was studied by empirical force field calculations. Estimation of heats of formation of four strain-minimum conformers revealed, in addition to the known conformer CT, the possible presence of another comparably stable conformer, CC. The ring inversion barrier between the CT enantiomers was estimated at AH* = 14.17 kcal/mol. We suggest that the newly recognized CC conformer is responsible for the biosynthesis of hirsutanoid and bicyclohumulenone and a transannular cyclization reaction to yield bicyclohumuladiol.
Humulene is one of the fundamental compounds in sesquiterpene biosynthesis. Farnesyl pyrophosphate undergoes head-to-tail intramolecular cyclization to produce humulenic and germacrenic cations and from both of them versatile cyclic sesquiterpene skeletons are derived' (Figure 1). In 1966, Allen and Rogers2 found that a tricyclic br~mohydrin,~ derived from humulene by treatment with N B S in aqueous acetone, had a conformation quite similar to that of humulene-AgN03 complex, by means of X-ray crystallographic analysis4 (Figure 2 ) . Since then, the conformational similarity between a reactant and the product has been discussed in several papers5 describing humulene, germacrene, and other medium-sized cycloolefin chemistry. In the course of studies on the biogenetic-like synthesis of illudoids,6 we were interested in the relationship among humulene conformations, its transannular in vitro reactions, and biosyntheses of humulene-derived sesquiterpenes. Roberts and his co-workers studied the conformation of humulene by means of N M R spectroscopy and estimated the ring flipping barrier. They were, however, unable to obtain information about the actual shapes of the stable conformation^.^ In the absence of any experimental clue to determine the predominant conformations of this apparently flexible molecule, we resorted to molecular mechanics calculations* to assess the relative stabilities of its conformers. The ring inversion barrier was also calculated by the same method. Among the many empirical force fields being available,I0 the 0002-7863/80/1502-3208$01 .OO/O
most popular program, Allinger's "MMI",' work.
I
was used in this
Results and Discussion Inspection of the molecular model shows that, in the 11membered ring containing three endocyclic trans double bonds, the planes of the latter should be almost perpendicular to the plane of the ring. Thus the number of stable conformers is limited to the number of combinations of the directions of the three double bonds. Four stable conformations, CT, CC, TT, and TC,I* then can readily be envisaged (Figure 3). The CT form appears in the crystalline silver nitrate ~ o m p l e xNoting .~ the chirality of three double bonds, eight possible stable conformers are RSR-CT, RRR-CC, RRS-TC, RSS-TT,I3 and the four enantiomers. Each of them can be correlated with seven other conformers through single or multiple double bond plane rotation. The correlation diagram of the eight conformers is depicted in Figure 4.14Calculations of heats of formation for the conformers a t the corners and barriers of 12 interconversion processes along each edge of this cubeI5 will cover the energy surface for the conformational change of humulene. 1. Stable Conformers. Energy minimizations of the four principal conformers were successfully achieved to give their precise geometries and heats of formation. Table I summarizes heats of formation and dihedral angles of the fully relaxed basic conformers and their perspective stereodrawings are depicted 0 1980 American Chemical Society
